Ovarian Cancer: The Interplay of Lifestyle and Genes

Ovarian cancer is a highly lethal disease that is mostly diagnosed at an advanced stage. In Europe, only 36% of women with ovarian cancer can expect to survive 5 years. While our knowledge of ovarian cancer has changed substantially throughout the years, our understanding of its etiology still lacks far behind and there clearly remains a need to better understand this disease. Here, a collection of epidemiological studies is presented that were designed to determine etiological factors for ovarian cancer, as well as to identify and characterize complex interactions that influence ovarian cancer susceptibility. As such, the research described in this thesis can be divided in three major parts. First, we determine the association between environmental (i.e., non-genetic) factors and ovarian cancer risk. Secondly, we investigate which genetic variations influence ovarian cancer susceptibility. Finally, we document our efforts to identify, characterize and interpret complex gene-gene and gene-environment interactions. The studies described in this thesis are performed within the European Prospective Investigation into Cancer and Nutrition (EPIC), the Netherlands Cohort Study on diet and cancer (NLCS), and/or the Nijmegen Ovarian Cancer Study (NOCS). The studies described in Part 1 of this thesis, show a protective effect of full-term pregnancies, oral contraceptive use, hysterectomy, and younger age at natural menopause. In contrast, risk of ovarian cancer was higher among women with a longer menstrual lifespan and among women who had more than three miscarriages. We did not find any associations with age at first birth, age at menarche, induced abortions and coffee and tea consumption. Part 2 of this thesis focuses on single nucleotide polymorphisms (SNPs) in association with ovarian cancer risk. We identified and validated a novel association between a missense mutation on chromosome 3q25. We also found evidence of associations between SNPs in the pleiotropic region TERT-CLPTM1L (5p15.33) and serous ovarian cancer risk. Furthermore, we show that SNPs in MYC and in the 9p22 region are differentially associated with type II ovarian cancer, while SNPs in CCND2 are associated with type I ovarian cancer only. Altogether, our findings suggest a role for genes in the DNA repair, cell cycle, and oncogenic pathways in ovarian carcinogenesis. Part 3 concentrates on complex high-order interactions between SNPs themselves and between SNPs and environmental factors. We first describe a strategy to detect and characterize gene-environment interactions, and subsequently identify and validate high-order SNP-smoking and SNP-SNP interactions in association with ovarian cancer. The results of this thesis show that ovarian cancer is influenced by a very complex interplay of lifestyle-related factors and genetic variations.